Using the green beam retro reflected by ETMY leaking into IR path, we found that the IR QPD path was off in YAW. I don't understand why/how, nobody touched this path since HIFO days, and this path worked in HIFO.
We fixed the problem by turning the pico mirror in the IR QPD path. Just tweek one mirror, and both of the QPDs were fixed.
Details:
There are many many many ghost beams because the reflectivity of IR HR for green is lower than that of IR AR, so we used foils to block ghost beams. It is easy to see the correct beam on QPDB, but was impossible to see the correct one on QPDA as the IR 50:50 acts as a good AR for green. Therefore, for QPDA, we looked at one specific ghost beam and used Kiwamu's calculation to set the ghost beam position right. At first both of the QPDs were off in the same direction by about the same amount. We twisted the TMS by giving it a bias offset and were able to bring the beam on QPDs to correct position. We removed the bias, tweaked one pico mirror until the beam positions were good. Unfortunately the green thing is so dim we couldn't see anything on the IR QPDs.
I noticed that one of the thicker cables running from the TMS ISC table to the top mass was half coming off of a J-clip on a post on the TMS ISC table on its own. I pushed the cable back in, and this made a large impact on the PIT of TMS such that we needed to rebalance.
There is some tension in the cable and this might be good enough to force the cable in the long run to creep out of the clip. Since the cable is stiff, once it changes its position there is an impact on the TMS balance. This might be the cause of mystery alignment jumps in OAT and HIFO days (i.e. cable creeps out of J-clip and settles to a new position).
We need to tighten J-clip some more tomorrow.
Remaining tasks tomorrow:
Here are a few photos from Thursday work. Took photos of Green beam on the Telescope mirrors, as well as our "bolt weight" which has a new position.